hf.co/unsloth/Qwen3.8-27B-GGUF:UD-IQ1_M
OLLAMA GGUFqwen35 · 27.3B · unknown
Aug 21, 2026 · AMD RYZEN AI MAX+ 395
hf.co/empero-ai/Qwen3.8-4B-Distill-GGUF:Q4_K_M
OLLAMA GGUFqwen35 · 4.33B · unknown
Aug 28, 2026 · Apple M4
Global Score
68 vs 66
Hardware Fit
69 vs 100
Quality Score
67 vs 52
Hardware
hf.co/unsloth/Qwen… hf.co/empero-ai/Qw…
MachineAZW GTR ProMacBook Pro
CPUAMD RYZEN AI MAX+ 395Apple M4
Cores3210
RAM125 GB16 GB
GPURadeon 8060SApple M4
OSUbuntu 24.04.4 LTSmacOS 26.6.1
Archx64arm64
Power Modelow-powerbalanced
Performance
hf.co/unsloth/Qwen… hf.co/empero-ai/Qw…
Tokens/sec21.126.5
First chunk571 ms447 ms
TTFT571 ms447 ms
Load time4.9 s3.3 s
Memory usage22.7 GB4.8 GB
Memory %18%30%
HW Fit Score Breakdown
hf.co/unsloth/Qwen3.8-2…
Speed
20/50
TTFT
19/20
Memory
30/30
hf.co/empero-ai/Qwen3.8…
Speed
50/50
TTFT
20/20
Memory
30/30
Quality
hf.co/unsloth/Qwen3.8-2…
Reasoning
12/20
Coding
10/20
Instruction
10/20
Structured
15/15
Math
10/15
Multilingual
10/10
Reasoning: Adequate Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Adequate Multilingual: Strong
hf.co/empero-ai/Qwen3.8…
Reasoning
18/20
Coding
14/20
Instruction
4/20
Structured
1/15
Math
15/15
Multilingual
0/10
Reasoning: Strong Coding: Adequate Instruction Following: Poor Structured Output: Poor Math: Strong Multilingual: Poor
Run yours and compare
$
npm install -g metrillm@latest$
metrillmRequires Node 20+ and Ollama or LM Studio running
Or run without installing: npx metrillm@latest